JP2023509892A - Temperature control system and method for semiconductor single crystal growth - Google Patents

Abstract

The present application discloses a temperature control system and method for semiconductor single crystal growth. The temperature control system includes an image acquisition device that determines the width of the ridgeline at the interface by capturing an image of the ridgeline of the crystal rod growing at the solid-liquid interface, a heating device that heats the crucible, and the heating device. a temperature control device for controlling the heating power of the heating device, wherein the temperature control device performs power control for the heating device based on the width of the ridge line. The temperature control system and method according to the present invention can significantly reduce defects in growing semiconductor single crystals, reduce manufacturing costs, and improve manufacturing efficiency.

Description

The present application relates to semiconductor single crystal growth techniques, and more particularly to temperature control systems and methods for semiconductor single crystal growth.

The Czochralski method is the main method for growing semiconductor single crystals. FIG. 1 is a schematic diagram of a typical Czochralski single crystal furnace, which mainly includes a crucible, a heating unit, a lifting rope, an observation window, a crystal melt, and so on. Taking single crystal silicon as an example, the process of growing a single crystal silicon rod is as follows: In the Czochralski single crystal furnace of FIG. Then, the heating unit controls the thermal electric field, and the lifting rope is used to rotate and slowly pull up the seed crystal to grow a semiconductor single crystal rod whose crystal orientation is consistent with that of the seed crystal. Here, the crystal directions are generally <100>, <110>, and <111> directions, and control of the thermal electric field is very important for the growth of semiconductor single crystals.

As semiconductor devices are getting smaller and smaller, the requirements for the quality of semiconductor wafers are also getting higher and higher, especially the requirements for wafer surface defects, flatness and surface metal impurities are continuously increasing. Therefore, controlling the axial temperature gradient at the solid-liquid interface to an appropriate range becomes increasingly important, because it is closely linked to the quality of the growing crystal rod. In recent years, it has been found that the axial temperature gradient at the solid-liquid interface is related to the width of the ridgeline of the growing crystal bar during the growth of a semiconductor single crystal by the Czochralski method.

By way of illustration, FIG. 2 shows a cross-sectional view of a crystal bar in the <100> crystal orientation. As can be seen from Fig. 2, the four ridges are in the <110> direction of the crystal. A relationship of '1' (L. Stockmeier et al. J. Cryst. Growth 515 26 (2019)) is given between the axial temperature gradient. Also, for the ridgeline of the crystal rod growing in other crystallographic directions, the ridgeline of the growing crystal rod is in a different position and orientation, but there is also correspondence between the width of the ridgeline and the axial temperature gradient at the solid-liquid interface. There is a relationship that However, until now, the axial temperature gradient at the solid-liquid interface cannot be accurately determined and effectively controlled.

The present invention provides a temperature control system and method for semiconductor single crystal growth to solve the above problems. Therefore, the present invention adopts the following technical solutions.

A temperature control system for semiconductor single crystal growth, said system comprising an image acquisition device for determining the width of said ridge at said interface by taking an image of the ridge of a crystal rod growing at said solid-liquid interface. a heating device for heating a crucible; and a temperature control device for controlling heating power of the heating device, wherein the temperature control device controls the heating device based on the width of the ridge line.

A temperature control method for growing a semiconductor single crystal, comprising the steps of capturing an image of a ridgeline of a crystal bar growing at a solid-liquid interface with an image acquisition device, and determining the width of the ridgeline based on the captured image. and, heating the crucible based on the width of the ridge.

1 is a schematic diagram of a typical Czochralski single crystal furnace; FIG. Figure 2 shows a cross-sectional view of a crystal bar in the <100> crystal orientation. 2D image of a growing semiconductor single crystal. Fig. 3(a) is a 2D image of the solid-liquid interface extracted from the image of Fig. 3(a); FIG. 4 shows a principle diagram of predicting the corresponding position and width of a ridgeline by processing a 2D image of a solid-liquid interface according to an embodiment of the present invention; FIG. 4 shows a curve diagram of the relationship between the width of the <100> crystallographic direction ridgeline and the axial temperature gradient at the interface, according to an embodiment of the present invention. FIG. 2 shows a schematic diagram of a heating device for controlling an axial temperature gradient at a solid-liquid interface, according to an embodiment of the present invention; FIG. 5 is an illustration of stepped heating power applied to a heating element for a stepped pre-gap heating method, in accordance with an embodiment of the present invention;

Embodiments of the invention will now be described in greater detail with reference to the drawings, which are intended to be illustrative rather than limiting.

According to an embodiment of the present invention, a photograph is first taken to determine the width of the ridgeline of the growing crystal. Image acquisition by the camera is set to be externally triggered. Add an IR bandpass filter external to the camera lens to avoid excessive brightness or interference from external light sources. FIG. 3(a) is a 2D image of a growing semiconductor single crystal captured by a camera. Preferably, the camera is placed in an observation window to photograph the growing crystal bar.

In one embodiment, the camera may employ a dual line scan camera, or any other high resolution camera. In another embodiment, if the diameter of the growing crystal rod can be obtained, a single line scan camera can be adopted, because then the diameter of the growing crystal rod and the real-time rotation speed of the crystal rod can be obtained. can be used to determine the width of the edge. As can be seen from FIG. 3(a), the seed crystal is continuously rotated and pulled up to produce a semiconductor crystal rod from the surface of the melt, and the positions indicated by the arrows are the corresponding positions of the ridges.

FIG. 3(b) is the corresponding 2D image at the solid-liquid interface extracted from the image of FIG. 3(a). As shown in FIG. 3(b), the white part at the bottom is an image of the solid-liquid interface, which is almost arc-shaped due to the viewing angle of the photograph. have different curvatures. Also, to simplify processing and calculations, the 2D image of the solid-liquid interface may be cropped even shorter, as shown in FIG. be.

After acquiring a 2D image of the solid-liquid interface, the position and width of the ridgeline is determined by processing the image as follows. Specifically, determining the position and width of the ridge from the image includes steps of extracting the interface edge curve, calculating the curvature, and comparing the curvature change with a threshold. FIG. 4 shows a principle diagram of predicting the corresponding position and width of a ridgeline by processing a 2D image of a solid-liquid interface according to an embodiment of the present invention.

For convenience of explanation, the upper part of FIG. 4 shows a 2D image of the solid-liquid interface of FIG. 3(b). Multiple search methods are used for the 2D image to find interface edge curves, and curvatures are calculated for the found interface edge curves. As shown in FIG. 4, the curvature of the edge curve is obtained by fitting the initial edge curve with a polynomial and taking the first derivative of the fitted curve, and the curve fitted with the polynomial is the initial curve and In order to be in good agreement, the drawings are basically collinear and overlapped, but this is by way of illustration and not limitation.

Also, as can be seen from FIG. 4, the curvature of the edge curve has a small value at the non-edge position and a large value at the corresponding position of the edge, because the curvature at the edge is large. Then, the curvature change curve of the edge curve is obtained by determining the curvature change and obtaining the second derivative with respect to the edge curve, as shown in FIG. Then, based on the curvature change of the edge curve, define a threshold line, determine the peak position of the curvature change curve and the intersection position with the threshold line, and the peak position of the curvature change curve is the determined ridge position. , and the two intersection locations of the threshold line is the width w of the determined edge.

The width of the ridge line does not remain unchanged. As the crystal bar grows, the heat dissipation area becomes larger and larger, the temperature gradient at the solid-liquid interface becomes higher and higher, and as a result, the width of the ridge line becomes smaller and smaller. Become. Therefore, according to the above, after determining the width of the ridgeline, in order to maintain the width of the ridgeline within an appropriate range, based on the predetermined width of the ridgeline, the theoretically established relationship is used to determine the solid-liquid interface to control the axial temperature gradient at By way of illustration, FIG. 5 shows a curve diagram of the relationship between the width of the <100> crystallographic direction ridges and the axial temperature gradient at the interface. As can be seen from FIG. 5, the temperature gradient at the solid-liquid interface must be maintained at 60-90 K/cm in order to stably maintain the ridge width within the range of 2 mm-6 mm. Hereinafter, a temperature control apparatus and method according to embodiments of the present invention will be described.

A conventional heating device comprises a main heating member and a bottom heating member, the main heating member being provided on the side wall of the crucible to heat the crucible from the side wall, and to prevent condensation on the liquid surface, the solid-liquid interface is heated. Heat across. Conventional heating devices do not independently control the heating of the main heating element and the bottom heating element. The present invention spans the solid-liquid interface and heats both sides of the interface simultaneously, with no significant change to the axial temperature gradient at the interface, and the bottom heating member is far from the interface, so the axial temperature at the interface is reduced. Consider the situation where the change to the temperature gradient is more pronounced.

FIG. 6 shows a schematic diagram of a heating device for controlling the temperature gradient at the solid-liquid interface, according to one embodiment of the present invention, in which to effectively control the axial temperature gradient at the solid-liquid interface, the required , the main heating element and the bottom heating element are independently controlled differently. Heat is preferably generated by resistive heating and radiated in a radiant manner into the crucible and further conducted from the crucible to the melt to heat the melt. Since the bottom heating member is far from the liquid surface, its power is relatively high, preferably 20-25% of the power of the melt, and the main heating member is close to the liquid surface, so the power is either too high or too low. If it is too high, the influence on the liquid level is large, so the power is controlled to 3 to 10%.

In FIG. 6, the heating elements at the sidewalls transfer heat in the direction of arrow A to heat the melt in the crucible from the sidewalls, and the heating elements at the bottom transfer heat in the direction of arrow B. , heats the melt in the crucible from the bottom. Then, part of the heat is transferred to the surface of the melt by convection and diffusion of the melt, and is taken away by the argon gas on the surface (C), and part of the heat is absorbed by the solid-liquid interface through phase change absorption, After being transmitted to the crystal body (D), it is emitted from the surface of the crystal body into the argon gas (E). The heating element at the bottom is located on the melt side of the solid-liquid interface, so that the temperature gradient at the solid-liquid interface can be modified more effectively.

For example, if the width of the ridge increases, it means that the axial temperature gradient at the solid-liquid interface has decreased, in which case the power of the bottom heater is increased and the power of the side heater is decreased. This moves the maximum temperature of the crucible down in the melt, lengthening the path of the buoyant vortices and increasing the axial temperature gradient at the solid-liquid interface. Conversely, when the width of the ridge becomes smaller, the axial temperature gradient at the solid-liquid interface becomes larger, which means that heat dissipation is improved. Increasing the power of the side heater moves the maximum crucible wall temperature up in the melt, shortening the path of the buoyant vortices and reducing the axial temperature gradient at the solid-liquid interface.

Also, conventional heating devices are typically supplied with constantly varying power and require a long time to transfer heat to the solid-liquid interface. In order to increase the thermal equilibrium rate and not lose the single crystal configuration, the present invention utilizes a stepped pre-gap heating method, which is different from conventional heating methods, to achieve thermal equilibrium more quickly. The step-type pre-gap heating method adopts a gradual increase of alternating heating power of increase-decrease-increase based on the rate of increase of the heating power, or alternately of decrease-increase-decrease based on the rate of decrease of the heating power. A gradual reduction of the mold heating power is adopted.

Hereinafter, the stepwise pre-gap heating method according to the present invention will be specifically described with reference to FIG. FIG. 7 shows a stepped increase in heating power and a stepped decrease in heating power when using the stepped pre-gap heating method. The stepwise temperature drop in FIG. 7 will be described below as an example. When the power of the heating device is reduced from 84 KW to 72 KW, first determine the power change rate and the number of steps. should also be selected appropriately.

By way of example, in FIG. 7 the power change rate is determined to be 1 KW/MIN and the stepped ramp down of the heating power is divided into 24 steps. Here, FIG. The abscissa is the number of steps and the ordinate is the heating power. Then, using the oblique line of the power change speed as a reference line, for example, the power value of the tenth step is estimated, and the estimated value is set as the fixed value of the first to third steps. After that, the baseline value of the sixth step is determined to be the fixed value of the fourth to sixth steps.

Next, the value of the 16th step is estimated using the reference line as the fixed values of the 6th to 9th steps. Thus, by determining the value of each subsequent step, we obtain a total of 24 steps of stepped down heating power. Regarding the stepwise increase of the heating power, the determination of the power value in each step is similar to the stepwise decrease, so the description is omitted here.

As described above, the present invention mainly determines and controls the axial temperature gradient at the solid-liquid interface by observing the width of the ridgeline of the growing semiconductor single crystal in real time, thereby achieving defect-free To achieve the purpose of manufacturing a semiconductor single crystal of The semiconductor single crystal manufactured by the system and method of the present invention has no crystal defects, and the semiconductor chip factory does not suffer from the loss of yield due to crystal defects on the surface of the silicon wafer during the device manufacturing process. Also, the system and method of the present invention can further improve manufacturing efficiency and reduce manufacturing costs.

The above examples are merely preferred implementations of the present invention, and are not limitations on the technical solution of the present invention. This invention should be considered to be within the scope of the claims of this patent.

For example, if the width of the ridgeline increases, it means that the axial temperature gradient at the solid-liquid interface has decreased, increasing the power of the heater at the bottom and increasing the power of the heater at the side wall . Reducing the power moves the maximum crucible temperature down in the melt, lengthening the path of the buoyant vortices and increasing the axial temperature gradient at the solid-liquid interface. Conversely, when the width of the ridge becomes smaller, the axial temperature gradient at the solid-liquid interface becomes larger, which means better heat dissipation, and at this time the power of the heating device at the bottom is reduced. , increasing the power of the heating device at the sidewall moves the maximum crucible wall temperature up in the melt, shortens the path of the buoyant vortices, and reduces the axial temperature gradient at the solid-liquid interface. .

Next, the value of the 16th step is estimated using the reference line as the fixed value of the 7th to 9th steps. Thus, by determining the value of each subsequent step, we obtain a total of 24 steps of stepped down heating power. Regarding the stepwise increase of the heating power, the determination of the power value in each step is similar to the stepwise decrease, so the description is omitted here.

Claims (14)

an image acquisition device for determining the width of the ridgeline at the solid-liquid interface by capturing an image of the ridgeline of the crystal rod growing at the solid-liquid interface;
a heating device for heating the crucible;
A temperature control system for semiconductor single crystal growth, comprising: a temperature control device for controlling heating power of the heating device,
The system according to claim 1, wherein the temperature control device performs power control for the heating device based on the width of the ridge line. 2. The system of claim 1, wherein the growth direction of the crystal bar is 100-direction, 110-direction or 111-direction. The heating device comprises a plurality of heating members respectively provided on the side wall and the bottom of the crucible, the heating members on the side walls heating the crucible from the side walls, and the heating members on the bottom heating the crucible from the bottom. 2. The system of claim 1, wherein: When the width of the ridge line is smaller than the predetermined range, the heating power of the heating member at the side wall is increased and the heating power of the heating member at the bottom is decreased, thereby reducing the axial temperature gradient at the solid-liquid interface. let
When the width of the ridgeline is larger than the predetermined range, the heating power of the heating member at the side wall is decreased and the heating power of the heating member at the bottom is increased, thereby increasing the axial temperature gradient at the solid-liquid interface. 4. The system of claim 3, wherein: 10. A system according to any one of the preceding claims, wherein the heating power of the heating element is increased or decreased using a stepwise pre-gap heating method. The step-type pre-gap heating method adopts a gradual increase of alternating heating power of increase-decrease-increase based on the rate of increase of the heating power, or decrease-increase-decrease based on the rate of decrease of the heating power. 6. A system according to claim 5, characterized in that it employs a gradual reduction of alternating heating power. A system according to any preceding claim, wherein the image acquisition device is a dual line scan camera or a single line scan camera at an observation window. 2. The system of claim 1, wherein determining the width of the ridge from the image comprises extracting the interface edge curve, calculating the curvature, and comparing the curvature change to a threshold. an image acquisition device capturing an image of a ridge line of a crystal rod growing at the solid-liquid interface;
determining the width of the ridge line based on the captured image, the temperature control method for semiconductor single crystal growth comprising:
A method, wherein the crucible is heated based on the width of the ridge. When the width of the ridgeline is smaller than a predetermined range, increasing the heating power at the side wall of the crucible and decreasing the heating power at the bottom of the crucible reduces the axial temperature gradient at the solid-liquid interface,
When the width of the ridge line is larger than a predetermined range, the heating power at the side wall of the crucible is decreased and the heating power at the bottom of the crucible is increased, thereby increasing the axial temperature gradient at the solid-liquid interface. 10. The method of claim 9, characterized by: 11. The method of claim 10, wherein the heating power is increased or decreased using a stepwise pre-gap heating method. The step-type pre-gap heating method adopts a gradual increase of alternating heating power of increase-decrease-increase based on the rate of increase of the heating power, or decrease-increase-decrease based on the rate of decrease of the heating power. 12. A method according to claim 11, characterized by employing a gradual reduction of alternating heating power. 10. The method of claim 9, wherein determining the width of the ridge from the image comprises extracting the interface edge curve, calculating the curvature, and comparing the curvature change to a threshold. 10. The method of claim 9, wherein the growth direction of the crystal bar is 100-direction, 110-direction or 111-direction.

Images